Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Yeah, I've never used Postscript except as a document format created by LaTeX :-)

There was a language called "V" a while back, different than a more recent language called V. It was basically a Forth where quoting was done with square brackets. This replaced the colon-semi notation for defining words, and it was also nice for nested data structures. This language seems to have fallen off the web though.

You mentioned FExprs. I never looked at Mock Lisp, and it sounds like Gosling doesn't think I should! However, I'm sure you're aware of Kernel. I think of Scheme as the "prettiest programming language I don't want to use", and I think the vau stuff in Kernel makes it even prettier. (But I still don't want to use it.)

For homoiconicity, I've also considered something like Tcl or Rebol/Red. The latter two blur the lines between lexer and parser in a way that I'd like to learn more about.

But really, I always come back to wanting static typing. Both for compile time error checking, and to give the compiler a leg up in runtime performance. Instead of using separate declarations like you see in Typed Racket and some others, I wonder if a Lisp with the addition of one "colon operator" to build typed-pairs would do it. Just one step down the slippery slope of infix syntax sugar. In the context of WebAssembly, something like this:

    (import (foo a:i32 b:f64):())
    (export (bar x:i64 y:f32):(i32 i32)
        (code goes here)
    )
Using colons to specify the types of the parameters and return result(s). It'd also be nice to have colon pairs like this for key:value in hash tables, or as cond:code in switch/conditional constructs.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: